Volume-associated hemodynamic variables for prediction of cardiac surgery-associated acute kidney injury

Abstract

Delayed diagnosis of acute kidney injury (AKI) is common because the changes in renal function markers often lag injury. We aimed to find optimal non-invasive hemodynamic variables for the prediction of postoperative AKI and AKI renal replacement therapy (RRT). The data were collected from 1,180 patients who underwent cardiac surgery in our hospital between March 2015 and Feb 2016. Postoperative central venous pressure (CVP), mean arterial pressure (MAP), heart rate, PaO2, and PaCO2 on ICU admission and daily fluid input and output (calculated as 24 h PFO) were monitored and compared between AKI vs. non-AKI and RRT vs non-RRT cases. The AKI and AKI-RRT incidences were 36.7% (n = 433) and 1.2% (n = 14). Low cardiac output syndromes (LCOSs) occurred significantly more in AKI and RRT than in non-AKI or non-RRT groups (13.2% vs. 3.9%, P < 0.01; 42.9% vs. 7.1%, P < 0.01). CVP on ICU admission was significantly higher in AKI and RRT than in non-AKI and non-RRT groups (11.5 vs. 9.0 mmHg, P < 0.01; 13.3 vs. 9.9 mmHg, P < 0.01). 24 h PFO in AKI and RRT cases were significantly higher than in non-AKI or non-RRT patients (1.6% vs. 0.9%, P < 0.01; 3.9% vs. 0.8%, P < 0.01). The areas under the ROC curves to predict postoperative AKI by CVP on ICU admission (> 11 mmHg) + LCOS + 24 h PFO (> 5%) and to predict AKI-RRT by CVP on ICU admission (> 13 mmHg) + LCOS + 24 h PFO (> 5%) were 0.763 and 0.886, respectively. The volume-associated hemodynamic variables, including CVP on ICU admission, LCOS, and 24 h PFO after surgery could predict postoperative AKI and AKI-RRT.